It becomes fetid if it's not constantly fed by a stream of fresh data at the expense of very valuable insights and correlations. 

It's Borges's The Library of Babel in an even more nightmarish form, if that's possible. This may account for a new kind of anorexia, in which the desire to communicate which had previously taken the form of a poem, short story or casual essay is slowly dying. The dialogue between literature and readers is like a stream or lake whose waters become fetid if they're not fed by new sources. The analogy is apt since flooding can wash out lakes, streams and rivers. It's either feast or famine, a total addictive indulgence with so much noise being created that individual voices are drowned out or a drought in which a human ability, like cursive writing, dies through attrition.

I am surprised that it seems like no one is working on curation and filtering solutions.

Everyone seems to think expensive and rigid complex event processing using existing data schemas are the only way to do it.  My personal view is dynamic tagging and scalable, schema-less data management, and lightweight search and filtering could easily clear the backlog of people wanting to do these types of projects.   I can't believe that there aren't hundreds, if not thousands of unfunded projects moping around because they can't get any time, money, or priority!

Yes, I think people are an important part of the solution. People WANT to help!